Psychodynamic therapy: a poorly defined concept with questionable evidence

ABSTRACT FROM: Leichsenring F, Luyten P, Hilsenroth M, et al. Psychodynamic therapy meets evidence-based medicine: a systematic review using updated criteria. Lancet Psychiatry 2015;2:648–60.

What is already known on this topic

Psychodynamic therapy (PDT) is widely practiced, but the empirical evidence for it is unclear.¹ As relatively few well-controlled studies exist, some authors resort to meta-analyses that include numerous poorly controlled and underpowered clinical trials in order to support their claim that PDT is efficacious.²

Methods of the study

After reviewing the distinction between superiority, non-inferiority and equivalency trials, Leichsenring and colleagues conducted a traditional literature search, which identified 64 randomised controlled trials (RCTs) examining the efficacy of PDT in common mental health disorders. The authors then proceeded to provide their own interpretation of the literature by concluding that PDT is as effective as cognitive–behavioural therapy (CBT). However, this conclusion is unjustified given the poor or unknown quality of the studies included in this review.

What does this paper add

• The general discussion about superiority, non-inferiority and equivalency trials was clear and accurate.

• This paper highlights problems in psychotherapy research. The authors provide a biased review of the literature, ignoring the quality of the data, and overstating the results. This article also reveals issues in the peer-review process of a high-level psychiatry journal.

Limitations

• The term PDT is poorly defined and appears to identify a variety of different interventions. The authors defined this intervention as an ‘umbrella concept for treatments that operate on an interpretive-supportive continuum’. It is difficult to imagine which forms of treatments do not meet this definition. CBT, for example, certainly does.

• The mechanism through which PDT is supposed to work remains unclear. The authors cited one study published in a book chapter, but none of the RCTs adequately examined treatment mediation.

• Treatments that ranged in length from 8 to 40 sessions were combined in the meta-analysis. This heterogeneity makes the results impossible to interpret.

• Attrition and dropout rates were not considered. Patients who benefit from treatment are more likely to stay in treatment, which may have biased the results.

• Therapy allegiance was not considered. This is particularly an issue for trials that compare PDT with CBT, because the trials comparing CBT with PDT are likely to have been conducted by PDT-oriented investigators.

• The placebo effect and common factors were not considered. No quality assessment of the individual trials was performed. Treatment integrity and adherence to the treatment protocol was not considered. The assessment methods of patients' symptomatology appear to have been inadequate.

• The 64 RCTs that were identified were generally of poor quality and underpowered. The CBT literature shows a very large number of high-quality RCTs. A review of meta-analyses examining the efficacy of CBT yielded 269 quantitative reviews.³ The vast majority of these studies showed clear superiority of CBT over other treatments and control conditions.

What next in research

• Rather than conducting yet another review of essentially the same poorly controlled studies from a different perspective, this field needs high-quality RCTs that clearly demonstrate the efficacy of PDT. For example, the Cochrane Collaboration's Tool is an instrument to quantify the risk of bias based on a number of criteria (eg, allocation concealment, blinding of participants and personnel, selective outcome reporting, etc).³ These trials will also need to test the hypothesised treatment mediation model. Without a clear understanding of the therapeutic process, it is difficult to distinguish them from other treatment modalities and to further improve the therapeutic strategies. In addition, therapy research has to identify patient characteristics that predict outcome.

Do these results change your practices and why?

No, this article does not provide any information that would change practice. Clinical practice should be based on empirical evidence. The empirical evidence for PDT is insufficient. The best evidence so far comes from high-quality RCTs comparing CBT and other treatments to adequate control groups.⁴ Based on this evidence, some countries have begun large-scale dissemination efforts, leading to substantial reductions in overall healthcare cost. These are the most promising strategies to change clinical practice.